Evaluation Of Health Status Of Lithium Battery Based On Internal Resistance Mode

With the popularization of lithium ion batteries in aerospace,new energy vehicles,civil power generation and other fields,ensuring the safety and reliability of lithium batteries under complex operating conditions has become a hot issue in the field of lithium battery research.State of Health(SOH)can directly represent the current State and service life of lithium battery,accurately estimate the SOH of lithium battery,and it is very important to evaluate the Health status of lithium battery.In order to solve the above problems,this paper has done the following research work for the establishment of high-precision lithium battery SOH prediction model:(1)based on the structure and working principle of lithium battery,the aging mechanism of lithium battery is expounded in depth,and the main reasons for the decline of SOH of lithium battery are analyzed;With the model 18650 lithium cobalt acid battery as the research object,the high-precision lithium battery test platform was built.From the perspective of testing the aging parameters of lithium battery,the aging characteristics test experiment of lithium battery is designed,which mainly includes the pretreatment cycle experiment,available capacity experiment,accelerated aging experiment and Hybrid Pulse Power Characteristic(HPPC)test experiment.The aging characteristic data of lithium battery were collected through experiments to establish a database for SOH estimation algorithm of lithium battery.(2)the advantages and disadvantages of electrochemical model,empirical model and equivalent circuit model were compared from three aspects of structural complexity,simulation accuracy and construction difficulty,and the Thevenin equivalent circuit model was selected as the research model in this paper.Based on the HPPC experimental data,the electrochemical mechanism in the discharge phase of lithium battery was analyzed,the parameters of Thevenin model were identified,and the accuracy was verified by Simscape simulation model.The influence of sampling period and current multiplier on the accuracy of Thevenin model was compared and analyzed.An improved HPPC experimental method was proposed under the conditions of 0.05 s sampling period and 1.5c charge-discharge current multiplier through experimental verification.The aging characteristic data of lithium battery were deeply studied,and the correlation degree between ohmic resistance,polarization resistance,cycle times and lithium battery SOH was quantitatively analyzed,so as to determine the characteristic input value estimated by lithium battery SOH.(3)BP neural network and support vector regression algorithm(SVR)were used to estimate the SOH of lithium battery.By comparing and analyzing the estimation errors of the two algorithms under different Numbers of training samples,it is found that the estimation errors of the SVR algorithm have a small fluctuation range,and the lithium battery SOH estimation supported by small sample data is more suitable for the establishment of highprecision lithium battery SOH prediction model.